Surrogate modelling (sometimes called black-box modelling, meta-modelling, or response surface modelling) is a way to build a simple mathematical model that can mimic a complex simulation. On the 3DEXPERIENCE platform, using SIMULIA apps like ABAQUS, Isight, or the Process Composer, this approach helps engineers save huge amounts of time.

Instead of running a full simulation for every design variation, you run a limited set of simulations once. Then, the surrogate model learns from those results and can instantly predict outcomes—like stress, deformation, or displacement—for new input values.

1. Design of Experiments (DOE): Pick different combinations of input parameters (e.g., length, thickness, or position).
2. Run Simulations: Use ABAQUS to run FEA simulations for each design point.
3. Build Surrogate Model: Fit a simpler model (e.g., polynomial regression, kriging, or neural networks) to the simulation results.
4. Prediction & Optimization: Use the surrogate to perform sensitivity analysis, optimization, or uncertainty quantification efficiently.

This process is often integrated through SIMULIA Process Composer App or SIMULIA Isight, which can work alongside ABAQUS to automate surrogate model creation and design optimization.

• Design Optimization: Quickly explore a large design space and find optimal parameters without needing to run thousands of physics simulations.
• Sensitivity Analysis: Understand how changes in inputs affect outputs.
• Uncertainty Quantification: Model and manage variability in material properties or boundary conditions.
• Real-time Simulations: Use the surrogate model in real-time systems where full simulations would be too slow.
• Multidisciplinary Analysis: Combine outputs from ABAQUS with other simulation tools in a streamlined optimization workflow.

Example: Formula Student Team

Imagine a Formula Student or SAE BAJA team testing a spaceframe chassis. Their goal is to reduce torsional stresses, but every design change requires hours of simulations. With so many load-bearing members in the chassis, trying different configurations could mean hundreds or even thousands of simulations, a huge drain on time and resources.

Here’s how surrogate modelling helps:

1. Set Parameters: Define key variables like the length or position of frame members.

2. Run DOE in ABAQUS: Perform a structured set of simulations to cover the design space.

3. Train the Surrogate Model: Use these results to build a fast predictive model.

Now, instead of waiting hours, the team can move sliders (for parameters like breadth or extrusion) and get updated results—such as mass or displacement—in just a few seconds.

This means time saved, money saved, and faster design decisions without compromising accuracy.

1.Surrogate models are low fidelity empirical models.

2. These are created bottoms up from simulation data

3. Capable of smoothing a noisy response

4. Extremely fast to evaluate!

5. Accurate.

Prerequisites of establishing a surrogate model:

1. Solve an FE Analysis with Geometric Parameters: 

Beam Analysis Using Geometric Parameters in FEA

Parameters for Length(Len), Breadth(BR) and Extrusion (ext)

2.Optimizing Simulations with DOE, ML, and Surrogate Modelling in SIMULIA

In the Optimization Process Composer app, a structured process is constructed to execute a Design of Experiments (DOE), generating a representative set of input-output pairs across the design space. The resulting numerical data feeds directly into a Machine Learning Model training pipeline, where surrogate modeling techniques such as Response Surface Modelling (RSM) or Universal Kriging (UK) are applied to approximate the underlying response behavior.

To enhance predictive accuracy, hyperparameter tuning is integrated into the workflow, employing optimization strategies to minimize the mean approximation error (e.g., Mean Squared Error or Root Mean Squared Error) of the surrogate models.

Once the surrogate model is trained and validated, it is made accessible within the Results Analytics App, enabling advanced post-processing, visualization, and sensitivity analysis using the approximated response surfaces.

Surrogate Modeling and Optimization Workflow

3. SIMULIA Results Analytics App

1. Data Extraction from ABAQUS Simulations

• Physics Results Analytics app allows you to visualize and extract results (e.g., displacements, stresses, temperatures) from simulation outputs.
• It supports efficient post-processing of large datasets—perfect for building surrogate models from a batch of simulation runs.

 2. Create and Analyze Datasets

• You can aggregate simulation results from multiple runs (DOE samples).
• Physics Results Analytics app helps in creating custom result metrics that can be used as outputs in the surrogate model.
• Supports filtering, grouping, and comparison across different parameter combinations.

 3. Link to Surrogate Modelling Tools

• Once the simulation data is extracted and prepared, PRA integrates seamlessly with tools like Isight, 3DEXPERIENCE Process Composer, or even external Python/ML tools.
• You can use this clean, organized dataset to train your response surface models, kriging, neural networks, or other surrogate types.

4. Perform Sensitivity & Correlation Analysis

• PRA includes built-in tools to analyze parameter sensitivities and correlations—helpful for feature selection before surrogate model training.
• This guides you in focusing on the most influential inputs, improving model accuracy and reducing complexity

5. Pre-process training of the model with ease 

 6. Integration into Design Exploration

• Designers can loop the analytics and surrogate models back into the design process for optimization, trade-off studies, or uncertainty quantification.
• PRA provides a user-friendly interface to track design decisions and results over time.

The Approximation model’s dashboard can be used this way:

Turning the breadth (BR) slider to 38.96

The FEA sensor output values update within 2 seconds with updated values.

Next, I will change the “ext” parameter while resetting the “BR” parameter.

The Mass and Displacement values update within 2 seconds.

All in all, we see that without actually going into CAD and FE apps on the platform, we get instant results of the FE analysis, saving time and eliminating repetitive workflows.

Key Benefits